SSRIC
Teaching Resources Depository
Exploring
the US Census
Eugene Turner, California State University, Northridge
Exercise
2 -- Analyzing Other Population Characteristics
Database:
USCOsp.por
In this exercise
you will use a few of the measures commonly used to describe population in more
detail.
- The
Sex Ratio
- Calculate
the sex ratio for California counties and for the state. Compare the state
total to that of the 58 counties. What reasons can you suggest for the large
differences between some of the counties?
a.
Under the Data menu select the Compute option.
(ie SexR = P005001 *
100 / P0050002)
- Compute the
sex ratio for other states and compare these to California. Why do you think
some states have higher and lower values?
- The
Location Quotient
- Compute the
location quotients for Class of Worker for all California counties and compare
these to the values for Los Angeles County in the module. What counties
have the greatest concentration of government workers and of self employed
persons?
- Compute location
quotients for the six civilian industry employment categories for California
counties. To calculate the location quotient, create a fixed variable of
the California proportion of the total employed in each industry category
and divide it into the county proportion of the total employed in each industry
category.
Item135
Percent Employed in Agriculture, Forestry, and Fisheries 1990
Item136 Percent Employed in Manufacturing 1990 and Retail Trade 1990
Item138 Percent Employed in Finance, Insurance, and Real Estate 1990
Item139 Percent Employed in Health Services 1990
Item140 Percent Employed in Public Administration 1990
Prepare a table
of counties (rows) versus employment categories (columns). Note which counties
have scores less than .3 or greater than 3 on any of the six employment categories.
Using a map of California and your knowledge of the geography of California,
explain as well as you can the reasons for any three of these unusually low
or high location quotients.
- The
Entropy Index
Compute the entropy index for California and its counties across five major
ethnic groups: non-Hispanic whites, blacks, American Indians (including Aleuts
and Eskimos), Asian and Pacific Islanders, and Hispanics.
Item005
Total population 1990
P0100001 Non-Hispanic white population
P0070002 Black population
P00703_05 American Indian-Eskimo-Aleut
population
P0070006 thru P0717_24 Asian and
Pacific Islander population
P0090002 thru P0090005 Hispanic
population
- First compute
a summary variable of the Asian and Hispanic populations. Then compute the
entropy index. In the equation below the white, black, American Indian,
Asian and Pacific Islander, and Hispanic populations are each divided by
the total population (Item5).
a.
To compute the entropy index select the Data menu and the Compute option.
H
= - ((P0100001 / Item5) * LN(P0100001 / Item5) +
(P0070002 / Item5)
* LN(P0070002 / Item5) +
(P00703_05 / Item5)
* LN(P00703_05 / Item5) +
(Asian / Item5)
* LN(Asian / Item5) +
(Hispanic / Item5)
* LN(Hispanic / Item5)) / 1.609
Note the leading
negative sign and the final division by 1.609. This number is the maximum
possible diversity score using the Log_{e} and five groups. It can
be determined by computing the H value for an equal proportion of an ethnic
population in each category. For example, if the five groups were evenly distributed
in a county, each group would have a proportion of 0.2 or 20 percent of the
population in the county. How do you think the value might change if you had
used census tracts instead of counties?
- Print out
the diversity index values along with the ethnic population percentages
for each of the counties. Which counties have the two highest and lowest
index values and why do you think these four counties are exceptional in
having such high or such low diversity? What groups dominate in counties
with low entropy? You have just identified the most and the least ethnically
diverse counties in California according to an appropriate statistical technique.
Are there other dimensions of diversity that you think should be incorporated
into its measurement statistically? Explain.
- Geographic
Association
- People often
assume that higher-income areas receive better health care. To test this
relationship, examine the association between the median household income
(Item79) with the infant death rate per thousand persons (Item52) for the
counties of the United States.
- Generate
a scattergram of these two variables to portray the strength and direction
of the association. Describe this relationship.
- Calculate
the Pearson product-moment correlation for the pair of variables. Is this
a significant relationship?
- Run a regression
on these two variables using household income as the independent variable
and infant death rate as the dependent variable.